Synthesis of Light-Emitting Conjugated Polymers for Applications in

Chem. Rev. 2009, 109, 897–1091

897

Synthesis of Light-Emitting Conjugated Polymers for Applications in Electroluminescent Devices Andrew C. Grimsdale,*,†,‡ Khai Leok Chan,†,§ Rainer E. Martin,| Pawel G. Jokisz,† and Andrew B. Holmes*,† School of Chemistry, Bio21 Institute, University of Melbourne, 30 Flemington Road, Victoria 3010, Australia; School of Materials Science and Engineering, Nanyang Technological University, Nanyang Avenue, Republic of Singapore 639798; Institute of Materials Research and Engineering (IMRE) and the Agency for Science, Technology and Research (A*STAR), 3 Research Link, Singapore 117602; and F. Hoffmann-La Roche Ltd., Pharmaceuticals Division, Discovery Chemistry, CH-4070 Basel, Switzerland Received January 5, 2007

Contents 1. Introduction 897 1.1. Background and Scope of the Review 897 1.2. Electroluminescence and LEDs 899 2. Poly(arylene vinylene)s 901 2.1. Poly(p-phenylene vinylene) (PPV) 901 2.2. Alkoxy-, Thioalkyl-, and Alkylamino-Substituted 905 PPVs 2.3. Alkyl- and Silyl-Substituted PPVs 917 2.4. Aryl-Substituted PPVs 921 2.5. PPVs with Electron-Withdrawing Substituents 926 2.6. Other Poly(arylene vinylene)s 931 2.7. Poly(heteroarylene vinylene)s 943 2.8. Poly(arylene ethynylene)s 955 3. Polyarylenes and Polyheteroarylenes 963 3.1. Poly(para-phenylene) (PPP) 963 3.2. Substituted PPPs 964 3.3. Ladder-type PPPs 966 3.4. Poly(9,9-dialkylfluorene)s and Related Polymers 967 3.5. Other Polyarylenes 977 3.6. Polythiophenes 978 3.7. Other Polyheteroarylenes 983 3.8. Polyarylene/Heteroarylene Copolymers 989 4. Other Luminescent Polymers 1001 4.1. Polyacetylenes 1001 4.2. Dendrimers and Hyperbranched Polymers 1002 4.3. Polymers with Luminescent Side-Chains 1010 4.4. Alternating Copolymers with Isolated 1017 Conjugated Groups 4.5. Block Copolymers 1032 5. Blends and Combinations of Polymers with Other 1039 Materials 5.1. Blends of Luminescent Polymers 1040 5.2. Blends of Luminescent and Nonluminescent 1046 Polymers 5.3. Blends with Nonpolymeric Materials 1048 5.4. Composites with Inorganic Materials 1050

6. Other Types of EL Devices Using Conjugated Polymers 6.1. Devices with Multiple Emissive Layers 6.2. Symmetrically Configured Alternating Current Light-Emitting Devices 6.3. Light-Emitting Electrochemical Cells 6.4. Electrogenerated Chemiluminescence Cells 6.5. Polymer Microcavities and Lasers 6.6. Integrated Polymer Devices and Other Devices 7. Limitations and Future Prospects of Electroluminescent Conjugated Polymers 7.1. Photo-oxidation of Conjugated Polymers 7.2. Polymer-Electrode Interactions 7.3. New LED Fabrication Methods 7.4. Commercial Materials and Device Lifetimes 8. Conclusion 9. List of Abbreviations Used 10. Acknowledgments 11. References

1051 1051 1052 1052 1054 1054 1055 1055 1055 1056 1057 1057 1058 1058 1059 1059

1. Introduction The award of the Nobel Prize for Chemistry in 2000 to Professors Alan Heeger, Alan MacDiarmid, and Hideki Shirakawa for their discovery that “doped” polyacetylene can conduct electricity changed the way that polymers were viewed and led to major advances in plastic electronics. Among these was the discovery of electroluminescent polymers and their application in light-emitting devices. These devices are now becoming commercially available, representing the final step in the advance of organic electronics from a laboratory curiosity to a part of everyday life. It is therefore appropriate to review the synthesis of such polymers and to provide a brief introduction to their applications. In this section we present a brief description of the historical background to the review and a concise account of the basic principles of electroluminescence and of light-emitting devices.

1.1. Background and Scope of the Review * Authors to whom correspondence should be addressed. E-mail: [email protected] and [email protected]. † University of Melbourne. Telephone: +61 3 83442344. Fax: +61 3 83442384. ‡ Nanyang Technological University. Telephone: +65 67906728. Fax: +65 67909081. § Current address: Institute of Materials Research and Engineering. | F. Hoffmann-La Roche Ltd.

Electroluminescence (EL) in conjugated polymers was first reported in poly(para-phenylene vinylene) (PPV, 1) in 1990.1 Since then considerable effort has been devoted to developing conjugated materials as the active units in light-emitting devices (LEDs) for use in display applications. A search of the ISI Science Citation Index database during the writing

10.1021/cr000013v CCC: $71.50  2009 American Chemical Society Published on Web 02/19/2009

898 Chemical Reviews, 2009, Vol. 109, No. 3

Andrew Grimsdale was born in Waiouru, New Zealand, in 1963 and graduated from the University of Auckland. He received his Ph.D. there under the supervision of Professor R. C. Cambie in 1990 for work on the synthesis of analogues of biologically active drimane sesquiterpenes. He then undertook postdoctoral work on photochromic and electroactive materials with Professor A. Pelter at the University of Wales, Swansea, and with Professor Andrew Holmes at the University of Cambridge on electroluminescent polymers. From 1999 to 2005 he was project leader in charge of research into conjugated polymers in the group of Professor Klaus Mu¨llen at the Max-Planck Institute for Polymer Research at Mainz. He rejoined the group of Professor Holmes in Melbourne in 2005 to work on materials for photovoltaic devices, before moving to Nanyang Technological University, in Singapore in November 2006. He is currently an Assistant Professor in the School of Materials Science and Engineering there, where his research interests are the synthesis of materials for optoelectronic applications and the formation of nanomaterials by selfassembly.

Khai Leok Chan was born in 1977 in Singapore. He obtained M.Sc. and Ph.D. degrees in Chemistry from University of Cambridge in 2002 and 2006 under the supervision of Professor Andrew B. Holmes. He returned to Singapore after graduation and joined DSO National Laboratories as a Member of the Technical Staff in the Advanced Materials Laboratory in 2006. He is currently a Research Scientist at the Institute of Materials Research and Engineering, under the Agency of Science, Technology and Research (A*-STAR), and his work focuses on materials and device designs for organic solar cells, thin film transistors, and light-emitting diodes.

of this review revealed that the original report of EL in PPV1 has been cited over 6200 times, impressively demonstrating the large degree of scientific interest. A further sign of the high level of interest is the fact that since our last review on electroluminescent polymers was published in 19982 (which has itself received over 1800 citations), the number of relevant publications has at least doubled. Interestingly, there has never been a fully comprehensive review of the synthesis and applications of electroluminescent polymers. Given that the number of publications on EL in conjugated polymers

Grimsdale et al.

Rainer Martin was born in Dornbirn, Austria, in 1970 and studied chemistry at the Eidgeno¨ssische Technische Hochschule (ETH) in Zu¨rich, where he received his diploma in 1995. In the group of Prof. F. Diederich, he worked on the synthesis and physical investigation of monodisperse poly(triacetylene) oligomers and obtained his Ph.D. in 1998. In September of the same year, he moved to the Melville Laboratory for Polymer Synthesis, Cambridge U.K., for postdoctoral research with Prof. A. Holmes on electroluminescent polymers supported by a fellowship from the Swiss National Science Foundation (SNF) and a Junior Research Fellowship awarded from Churchill College. He started his professional career at F. Hoffmann-La Roche in Basel in 2000 as Head of the Laboratory in Formulation Research and switched in May 2002 to Discovery Chemistry, working as a Senior Research Scientist in the department of Medicinal Chemistry.

Pawel Jokisz was born in 1975 in Warsaw, Poland. He completed his undergraduate Chemistry degree with Prof Andrew Holmes at the University of Melbourne in 2007 and will be starting his Ph.D. studies in the area of analytical chemistry early in 2009.

will soon be so great as to make a comprehensive review unmanageably large and that much future work will probably remain confidential for commercial reasons as these materials are now entering the marketplace, it is an appropriate time to present a comprehensive overview of the subject. Since 1998 a number of comprehensive reviews on organic electroluminescence have been published.3-14 A number of short reviews (in addition to those listed in our earlier work) have appeared covering the work of individual groups15-21 or of individual aspects of this field,22-30 including the physics of organic electroluminescence.31,32 There have also been some more general reviews on the synthesis, properties, and applications of conjugated polymers,33-36 including reviews of methods for patterning or orienting polymer films which are of potentially great importance for some device applications.37,38 This review is limited to polymeric organic materials. Conjugated oligomers as luminescent materials and as models

Synthesis of Light-Emitting Conjugated Polymers

Andrew Holmes was born in 1943 and is an Australian Research Council (ARC) Federation and Inaugural Victorian Endowment for Science Knowledge Innovation (VESKI) Fellow in the School of Chemistry/Bio21 Institute at the University of Melbourne. He also holds joint appointments at CSIRO Molecular and Health Technologies and Imperial College London from where he has been seconded on leave since 2004. After undergraduate education at the University of Melbourne, he completed a Ph.D. in 1971 at University College London (supervisor Professor F. Sondheimer). He worked as a Royal Society European postdoctoral fellow at E.T.H. Zu¨rich (Professor A. Eschenmoser) in 1971-2 and then held an academic appointment at the University of Cambridge from 1972 to 2004, where he was Professor of Organic and Polymer Chemistry and Director of the Melville Laboratory for Polymer Synthesis. His research interests are in the applications of synthesis to problems in materials science and biology, and he cofounded Cambridge Display Technology to exploit the applications of light-emitting polymers.

for conjugated polymers have recently been the subject of reviews by Martin and Diederich39 and by van Hutten and Hadziioannou40 and of a collection of monographs edited by Mu¨llen and Wegner,41 to which the reader is referred. In this review oligomers will be discussed only where they serve to illustrate a structure-property relationship, (e.g., shifts in emission spectra due to substituents), to determine a key polymer property (e.g., effective conjugation lengths), or to demonstrate a technique that may be applicable to polymers (e.g., orientation to obtain polarized emission). We have endeavored to cover comprehensively the literature up to the end of 2006 with the inclusion of a few later papers of particular importance which appeared during the final revision of the manuscript. We have concentrated on the academic literature and have not attempted to cover the patent literature, but we have cited a number of patents where these contain details of synthetic methodology not otherwise available. The study and application of electroluminescent materials is a multidisciplinary effort involving synthetic chemists, physicists, and device engineers. This review is written from a chemist′s viewpoint. Our aim is to show what types of structures have been made, how they can be made, and how their properties may be controlled by means of intelligent, rational synthetic design. We are not concerned here with details of optimization of device performance, but we shall cover aspects such as the use of charge-transporting layers and shall cite examples of where these are needed to optimize performance for individual polymers, as these factors are of relevance in designing a new polymer structure. Note also that device efficiencies will be quoted either as external quantum efficiencies (ratio of photons emitted as outcoupled light to the charge injected) expressed as a percentage or emission per unit current (cd/A) as given in the original literature. Usually these values are quoted for the maximum efficiency, which may not be the efficiency of the device at maximum brightness (quoted as cd/m2).

Chemical Reviews, 2009, Vol. 109, No. 3 899

Figure 1. Illustration of electroluminescence (EL) in a conjugated polymer material. In electroluminescence, holes are injected at the ITO anode (removal of electrons from the HOMO) to give positively charged species (radical cations), and electrons at the calcium cathode are injected into the LUMO to form negatively charged species (radical anions). The charged species are drawn to the oppositely charged electrode and migrate from polymer chain to the polymer chain by a hopping process as a result of the applied electric field. When the two opposite charges combine, charge annihilation leads to singlet and triplet excited states (excitons), of which the former can emit photons through fluorescence. The work functions of the two metal electrodes are shown on the figure.

Readers wishing to obtain full details of device performance are directed to the primary literatureswe will here cite numbers of the relative reported performance of different polymers only in order to illustrate possible ways to influence performance by means of structure.

1.2. Electroluminescence and LEDs Electroluminescence (EL) may be defined as the nonthermal generation of light upon the application of an electric field to a substrate; that is, it is the production of light by passing electricity through a material by means other than the simple incandescence seen in a standard filament light bulb. As shown in Figure 1, it results from recombination of charge carriers (holes and electrons) injected into a semiconductor in the presence of an external circuit. If these combine to give a singlet excited-state identical to that obtained in photoluminescence (PL) by excitation of an electron from the HOMO to the LUMO, then symmetrypermitted relaxation to the ground-state may occur with emission of a photon. The color of the emission obviously depends upon the size of the HOMO-LUMO energy gap, which for visible light (380-780 nm) corresponds to 1.5-3.2 eV. It should be noted that whereas chemists usually describe the energy associated with light in terms of its wavelength in nanometers (nm), physicists tend to refer to its energy in electronvolts (eV). In this review we will refer to the light emitted by the polymers using nanometers, and where we quote values from the literature in electronvolts, we will give their equivalent in nanometers. If the emitting material is oriented, e.g., by use of aligned liquid crystalline phases (the use of such phases in OLEDs has been reviewed by O’Neill and Kelly42), then the emitted light may be polarized, which is important in some applications, e.g. backlights for liquid crystalline displays (LCDs). The development of polarized electroluminescent devices for such applications has been reviewed by Grell and Bradley.43 EL has been known in inorganic materials for many years, and light-emitting devices (LEDs) using inorganic phosphors have been commercially available since the early 1960s. EL


Grimsdale et al.

Figure 2. Schematic drawing of a single-layer EL device. Double charge injection through application of a forward bias voltage results in formation of an electron-hole pair. The singlet excited state which is formed emits a photon by radiative decay to the ground state.

was first observed in organic materials in 1963 with anthracene crystals,44 but it was only with the development of thin films of efficient organic dye-based LEDs by Tang and VanSlyke45 and Saito and Tsutsui46 in the late 1980s that significant interest in organic EL devices emerged. The first commercial products using molecular organic LEDs appeared in 1997.9 Emissive devices using polymeric emitting layers were first made by Partridge in 1974.47,48 These used dyes incorporated into a polymeric charge-transporting matrix. The first LEDs using a conjugated polymer as emissive material were made in Cambridge in 1989 using PPV (1).1 Polymer LEDs (PLEDs) have now entered the commercial marketplace, with the release by Philips of an electric shaver with a monochromatic PLED display. A fuller discussion of the history of EL and of LEDs appears in the review article by Mitschke and Ba¨uerle.14 In a typical polymer LED (Figure 2), a thin film of an emissive polymer is sandwiched between two electrodes, at least one of which is transparent. Generally, commercially available indium tin oxide (ITO) on a glass or polymer substrate is chosen as the (transparent) anode, and the cathode consists of a vacuum-deposited metal layer. However, a variety of other materials have been used as anodes or cathodes. There are many factors that may contribute to the overall efficiency of an LED. A detailed description of the methods that have been used to optimize the efficiency of LEDs using conjugated polymers lies outside the scope of this review, but we will provide a general overview in this section.49 In general, the external efficiency, ηEL, of an OLED can be summarized by the following equation:

ηEL ) γηPLηCχS

(1)

where γ relates to the charge balance in the emissive layer, ηPL is the photoluminescence quantum efficiency of the emitter, ηC is the fraction of photons coupled out of the device, and χS is the singlet fraction. If charge mobilities are not balanced, recombination can occur close to one of the electrodes where traps and defects are likely to be prevalent. Therefore, control of the charge balance γ is an important factor in improving device efficiencies. Some further detailed discussion is also given for the specific cases of certain polymers, e.g. PPV (section 2.1) and MEH-PPV (section 2.2), as examples to illustrate the general principles. The efficiency of an LED depends greatly upon the choice of electrode materials, as the efficiency of

charge injection depends upon the size of the energy barrier between the work function of the electrode and the energy level of the HOMO (for the anode) or LUMO (for the cathode). Since balanced charge injection is required (emission requires a hole and an electron to combine), the energy barriers should also be of comparable sizes. As a result, in designing a new polymer for use in an LED, care must be taken to match the HOMO and LUMO energies as closely as possible to the desired electrode materials. These problems may be partially overcome by use of charge-transporting layers (sometimes called level-matching layers) whose energy levels are intermediate between those of the electrode and emissive material. For a discussion of the factors affecting charge injection, the reader is referred to the reviews by Greenham and Friend,50 Sheats et al.,24,51 and Schott.32 There are also some recent reviews on materials for hole-transport52 and electron-transport layers53 in LEDs. If an LED contains more than one layer, then the question of cross solubilities arises, as the polymer layers are deposited from solution, usually by spin-coating, and so it is essential that as each layer is deposited, it does not dissolve the preceding layer. Layers are sometimes rendered insoluble by cross-linking (thermal or photochemical) of suitable substituents in order to surmount this problem. Polymers that are insoluble, e.g. PPV (1), are usually deposited as soluble precursor polymers, which are then thermally or irradiatively converted to the final conjugated polymer. Though precursor routes are invaluable as a means for obtaining thin films of insoluble materials, the byproducts should not corrode the electrodes or adversely dope the emissive or other layerssa factor that must be considered in synthetic design. Electron-hole recombination of statistically generated spin states formed by double charge injection is predicted to lead to 25% of singlet states capable of spin-permitted radiative decay (fluorescence) and 75% of the relatively nonemissive and long-lived triplet states. Thus, the maximum EL efficiency obtainable would be predicted to be a quarter of the solid-state PL quantum efficiency,54 implying a maximum theoretical external efficiency for a polymer LED of 25%.55 The 3:1 singlet/triplet ratio has been experimentally confirmed in small molecule OLED devices.56 However, as was first reported by Rothberg et al.,57 the EL external efficiency for some polymer multilayer devices is apparently higher than 25% of the PL efficiency, although no research group has to our knowledge fabricated a PLED single-layer device exhibiting >25% external quantum efficiency. The value for the proportion of singlet states in conjugated polymer LEDs, and thus their maximum EL efficiency, is the subject of an ongoing debate in the physics literature, to which the reader is referred (see reviews by Ko¨hler58-61), but a number of experimental49,62-65 and theoretical66,67 results have shown that, at least for some polymers, the potential EL efficiency of a polymer device may be up to 50% of the PL efficiency, although this is still disputed.68 It has also been reported recently that blending cobalt-platinum nanowires with a fluorescent polymer increases the percentage of singlet excitons and thus the overall EL efficiency.69 Maximization of the solid-state PL quantum efficiency is clearly desirable for obtaining maximal EL efficiency from conjugated polymers. It is not yet apparent how this may be done for any given structural type, except by trial and error. However, there are some general rules. For example, the presence of known fluorescence quenching functionalities such as halide or carbonyl groups on the polymer chain as impurities



Scheme 1. Sulfonium Precursor Route to PPV (1): (a) Tetrahydrothiophene, MeOH, 65 °C; (b) NaOH, MeOH/H2O, 0 °C; (c) HCl; (d) Dialysis (water); (e) MeOH, 50 °C; (f) 220 °C, HCl(g)/Ar; (g) 180-300 °C, vacuum

(defects) formed during the reaction pathway should be avoided. As a result, the choice of synthetic pathway or reaction conditions may have a significant effect on the PL efficiency of the final polymer. The harvesting of energy from the triplet states is an increasingly important goal in research into OLEDs, as electrophosphorescent devices, i.e. LEDs utilizing emission from triplet states (phosphorescence), have potentially much higher efficiencies than those producing emission only from singlet states.70,71 Since phosphorescence from conjugated polymers, if present, is generally very weak, the “wasted” triplet energy can be harnessed by the incorporation of phosphorescent metal complexes into the polymers or the blending of such complexes with the polymers (see section 5.3). Considerable progress has been made so that phosphorescent OLEDs are already competitive with the conventional fluorescent OLEDs.72,73 Electrophosphorescent devices are likely to contribute strongly to the development of solid state lighting applications and to be present as the red component of all OLED/PLED displays. As can be seen, many factors must be taken into account in designing a new polymer and its synthesis. Extensive evidence of this follows in the remainder of this review.

2. Poly(arylene vinylene)s In this section we cover poly(arylene vinylene)s and related polymers such as poly(arylene ethynylene)s. Poly(arylene vinylene)s (PAVs) were the first class of conjugated polymers in which EL was demonstrated, and as a result, they form the most widely studied class of electroluminescent polymers. There are reviews on the methods for their synthesis by Scherf,26 by Denton and Lahti,74 and by Grigoras and Antonoaia75 and on structure-property relationships in PAVs by Shim and Jin.76

2.1. Poly(p-phenylene vinylene) (PPV) A more detailed account of the synthesis and processing of PPV than any of the other polymers is provided in this review, as the problems associated with the development of PPV and the solutions developed are illustrative of those seen for other materials. Thin films of PPV exhibit bright yellowgreen fluorescence with an emission maximum at 2.25 eV (λmax ) 551 nm) and a secondary peak at 2.4 eV (520 nm). Extrapolation from the spectra of monodisperse phenylene vinylene oligomers suggests that the effective conjugation length for solid-state fluorescence in PPV is between 8 and 10 repeat units.77 This correlates well with the value of 11

repeat units obtained from studies on substituted oligomers.78 (Longer values are found in solution, which is attributed to the stronger interchain interactions in the solid state and to the existence of torsional disorder in solution.79) Unsubstituted PPV is insoluble, intractable, and infusible. Any synthesis that results in the preparation of PPV directly from a monomer, therefore, produces only insoluble material which cannot be further processed. This is highly undesirable, as solution processing by spin-coating is particularly useful for the production of polymer EL devices. The use of a processible poly(p-xylylene) precursor polymer provides a convenient way around this dilemma. The synthesis and applications of poly(p-xylylene)s have been reviewed by Greiner.80 Introduced by Wessling and Zimmerman81,82 and later modified and optimized by several groups,83-88 the sulfonium precursor route to PPV has proven to be a versatile approach not only to PPV and its derivatives but also to a whole range of PPV-related copolymers. Scheme 1 outlines the standard preparation of PPV (1),87 using a bis-sulfonium salt monomer 2. Although other sulfonium salts, such as those derived from dimethyl sulfide or tetrahydrothiopyran, have been used; tetrahydrothiophene is normally chosen, as it represents a satisfactory compromise between expense on the one hand and suppression of undesired side reactions (which are more prone to occur with dimethylsulfonium groups83) during the subsequent conversion step on the other hand. Polymerization of a methanolic solution of the monomer is induced by addition of slightly less than 1 mol equiv, e.g. 0.9 mol equiv, of aqueous hydroxide base at 0-5 °C. If 1 equiv is used, then the resulting polymer solutions become brightly colored, owing to base-induced elimination of the sulfonium groups. Studies82,89 have shown that the presence of radical trapping agents such as oxygen can severely lower the molecular weight of the polymer, which would suggest that the polymerization proceeds Via a radical pathway (for a discussion of the mechanism of the Wessling polymerization, which is thought to involve the transiently formed quinomethide intermediate, the reader is referred to the review by Denton and Lahti74), and thus the reaction must be performed under an inert atmosphere. Alternatively, the polymerization can be performed in a dispersion of aqueous droplets in mineral oil using a high-speed stirrer or homogenizer.90 Polymer produced by this method shows similar properties to that prepared by the standard method. The molecular weight of the resulting precursor, polymer 3, cannot be determined directly by size exclusion chromatography (SEC)


Grimsdale et al.

Scheme 2. Sulfinyl Precursor Route to PPV (1): (a) RSH, NaOH; (b) H2O2, TeO2; (c) KOtBu; (d) 180-300 °C, vacuum

Scheme 3. Alternative Routes to PPV (1): (a) 500-700 °C, 0.01 mbar; (b) 580 °C, 10 Pa; (c) 200 °C, vacuum; (d) CH3CN, 5.5 V

Scheme 4. ROMP Routes to PPV (1): (a) [Mo(dNAr)(dCHCMe2Ph){OCMe(CF3)2}]; (b) Bu4NF; (c) HCl(g), 190 °C; (d) 105 °C; (e) 280 °C

due to the polyelectrolyte nature of the polymer, but treatment with refluxing methanol affords a neutral polymer 4, which on analysis by SEC shows a number-average molecular mass of ca. 100,000 g/mol, corresponding to about 750 repeating units, with a polydispersity of 5.7.91 Prasad and co-workers have prepared PPV by performing the polymerization inside reverse micelles.92 They were able to control the conjugation length by varying the micelle size, so that the PL maximum could be tuned between 493 and 505 nm. Although the purified, mainly aqueous, polymer solutions can be used directly, there are considerable advantages to isolating the polymer and redissolving it in methanol. Such solutions can be safely stored at -20 °C for several months. Solutions of the polyelectrolyte 3 in water,83 methanol,87 or cyclohexanone93 are highly viscous even at a concentration of only 1% (w/w) of polymer, which facilitates their

processing by spin-coating; however, using methanolic rather than aqueous solutions produces thin films of higher quality.87 Alternatively, the precursor polymer 3 has been cast with a precision doctor-blade machine which can coat areas of about 25 cm2.82,93 The sulfonium precursor polymer 3 is typically converted into PPV (1) by heating thin films at 220-250 °C for 12 h under vacuum (10-6 mbar).1,87 The luminescence efficiency of PPV produced by this method seems to be strongly dependent upon the conversion conditions. A detailed study on the mechanism of the elimination reaction found that the mechanism was complex and that reducing the heating rate and/or converting under a reducing atmosphere gave PPV with higher PL efficiencies.94 Friend and co-workers reported that the PL efficiency of PPV converted on ITO substrates was optimal at a conversion temperature of 205 °C.95 The

Synthesis of Light-Emitting Conjugated Polymers Scheme 5. Synthesis of Partially Conjugated PPV Copolymers: (a) Bu4NOH, MeOH; (b) MeOH, 50 °C; (c) 300 °C, vacuum; (d) 220 °C, HCl(g)/Ar

EL efficiency of single-layer devices increased with increasing conversion temperature, while that of double-layer devices using an electron-transporting layer decreased. A Korean group has reported that conversion at 230 °C for 5 min under nitrogen gave optimal EL efficiency in singlelayer LEDs.96 Using the bromide instead of the chloride salt enables the conversion to occur at temperatures as low as 100 °C.88,97 Replacing the chloride with a dodecylphenylsulfonate counterion also lowers the conversion temperature, so that complete conversion occurs rapidly at 115 °C.98,99 The properties of PPV prepared by elimination of sulfonium precursor polymers are found to depend upon the leaving group with an increase in intrachain order and a decrease in interchain order with increasing size of the leaving group.100,101 The neutral precursor polymer 4 can be converted to PPV by simultaneous heat and acid treatment. Halliday et al.102,103 reported that treatment of the precursor polymer 4 with base, followed by methanolysis, gave partially conjugated material 5 (Scheme 5). Subsequent conversion to the fully conjugated polymer produced PPV which was more highly ordered than standard PPV101 and which showed a red-shift in the absorption maximum.104 The properties of this material have been found to depend strongly on the experimental conditions applied, and oligomeric material may be responsible for the observed properties. Furthermore, the process is not easily amenable to EL device fabrication, as the conversion step tends to cause corrosion of the electrodes. A similar partially converted precursor polymer can be obtained by long-term storage of 3 in methanol at 0-5 °C. PPV prepared from this precursor is also reported to be more highly ordered than standard PPV.105 It has been found that traces of oxygen in the conversion atmosphere cause reduction in EL in the final PPV film due to the formation of carbonyl group defects,106-108 which can be suppressed by performing the elimination reaction in an atmosphere of nitrogen and hydrogen.106,107 PPV prepared under these conditions was found to have higher PL and EL quantum yields.94,109 Schwoerer and co-workers110-112 have found that the conversion temperature can be lowered to 160° by performing the conversion under an argon atmosphere. This enables LEDs to be fabricated on flexible polymer foil substrates. It has been reported113,114 that the elimination of


the sulfonium groups from the precursor polymer can be performed even at room temperature by use of protic acids followed by dedoping of the resulting polymer with aqueous sodium hydroxide. However, this method is unlikely to prove suitable for the production of PPV films for LED devices, owing to the problems involved in obtaining complete dedoping (acids are not compatible with most electrode materials) and also to the nucleophilic displacement of a proportion of the sulfonium groups by water, which will give rise to a reduced degree of conjugation in the resulting polymers. A study on the fluorescence of doped PPV films115 showed that doping with lithium or sulfuric acid gave red and blue-shifts, respectively, in the photoluminesence spectra but that the material properties of the films were adversely affected, making such polymer films unsuitable for use in electroluminescent devices. Elimination can also be induced by irradiation with UV light,116-118 laser light,119,120 or microwaves.121 Irradiation of PPV films with light leads to cross-linking and a consequent decrease in the luminescence of the film, which can be used as a means of patterning PPV films.122 Films of PPV prepared from Langmuir-Blodgett (LB) films of precursor polymer 3123 also show electroluminescence despite the fact that the PPV is contaminated by surfactant residues.124-129 Langmuir-Blodgett or selfassembled films of the sulfonium precursor 3 with dodecylphenylsulfonate counterions can be converted at temperatures as low as 80 °C.130-132 The fluorescence of PPV prepared from self-assembled films of the precursor and sulfonated polystyrene is reported to be blue-shifted with respect to the case of standard PPV, indicating a shorter conjugation length.133 Ordered LB films of PPV have been shown to emit circularly polarized PL.134 Other methods have also been exploited to prepare PPV for use in LEDs. A modified Wessling synthesis using a sulfinyl (or sulfonyl) precursor 6 has been reported by Vanderzande and co-workers (Scheme 2)135-142 which produces PPV with lower defect levels.143 A disadvantage of this method is that the monomer synthesis is more complicated though the initial low yield has since been greatly improved.144 This synthesis has been shown to proceed via a radical mechanism.145 The same group has recently reported that a dithiocarbamate precursor is easier to make, but the quality of the materials produced is not as good as that from the sulfinyl route.146 Galvin and co-workers have described a modified Wessling route to PPV using an organic soluble xanthate precursor polymer which gives better quality PPV films and higher device efficiencies than the standard route.147,148 PPV films have also been prepared by chemical vapor deposition of 1,4-bis(halomethyl)benzene 7149-151 or bishalocyclophane 8152 to give R-halo-precursor polymer film 9, which is converted to PPV at 200 °C under vacuum (Scheme 3). LEDs made from the chloro precursor show rather low (0.002%) EL efficiency,149,150 but films made from the bromo precursor have properties comparable to those of standard PPV.153 Use of the bromomethyl monomer 7 (X ) Br) also leads to the formation of more ethanediyl defects and thus a blue-shift in the PL of the resulting PPV.154 The use of UV light enables the conversion of the precursor to be performed at room temperature.155 PPV films can be grown selectively by CVD in patterns on surfaces treated with iron or other group VIII metal salts.156 PPV can also be produced electrochemically.157 Electropolymerization of a xylylene bisphosphonium salt 10 gives films of PPV (1)


whose absorption and emisssion spectra are blue-shifted with respect to those of materials produced by other methods, indicating that the electropolymerized material has a shorter conjugation length, probably due to incomplete elimination of phosphonium groups. A device fabricated using electropolymerized PPV emitted green (λmax 530 nm) light.158 It is reported that LEDs using electropolymerized PPV show lower EL efficiencies than those using standard PPV.159 A synthetically elegant approach to the preparation of PPV is the ring-opening metathesis polymerization (ROMP) of a bicyclo[2.2.2]octadiene monomer 11160 or of a silyloxy cyclophene 12,161 followed by thermal elimination of the resulting precursor polymers 13-15 (Scheme 4). A related approach using ROMP on a cyclophanediene has been used to prepare PPV derivatives with alternating cis/trans microstructure.162 As mentioned in the introduction, it has been clearly demonstrated that choosing the correct electrode combination is crucial to the efficient operation of polymer LEDs. In the case of PPV, the best early results (0.1%) with regard to device efficiency (photons emitted vs electrons injected) were achieved with indium tin oxide (ITO) as the hole-injecting electrode and calcium, a low work-function metal, as the cathode.163,164 However, a problem with converting the sulfonium precursor to PPV on ITO is that it leads to doping with indium chloride, which reduces the PL efficiency.165 It might be expected that this decrease in PL efficiency would also lower the EL efficiency, but Morgado et al. have reported that indium-doped PPV, prepared by doping the precursor polymer 3 with indium(III) chloride before conversion, showed a lower PL quantum efficiency but an increased EL efficiency.166 Other materials such as tin oxide have been explored in place of ITO as anode but without improvement in device performance.1,167 Use of a p-doped silicon anode offers the possibility of integrating PLEDs with microelectronic circuitry on a silicon chip.168 As a result of its reactivity (especially its moisture-sensitivity), calcium is not particularly suitable for use in commercial devices, although it can be encapsulated in aluminum. In order to construct devices which can be used outside a glovebox, less reactive metals such as aluminum are preferred. The efficiency of the first PPV LEDs using ITO and Al electrodes was only 0.01%, owing to poor electron injection.164 Introduction of a thin interfacial layer of aluminum oxide169 or a lithiumdoped polyarylene170 between the PPV and Al layers is reported to lead to a marked improvement in device efficiency due to charge carrier confinement. A double-layer cathode with a low work-function strontium layer sandwiched between air-stable indium and the PPV layer has also been found to give improved electron injection.171 It is reported that the emission spectrum from PPV varies with different metal cathodes, with the intensity of the peak at 550 nm being greater with lower work function metals.172 A similar increase in the relative intensity of the 550 nm emission peak has been observed upon increase of either the thickness of the PPV layer or the drive voltage in an ITO/PPV/Al device,173 or by increasing the temperature of the PPV conversion.174 The addition of an electron-transport layer between the PPV and the cathode has been found to considerably improve both the power efficiency and the stability of devices.54,164,165,175-191 Use of a hole-transporting layer on the other hand causes a decrease in efficiency, indicating that the main limiting factor in PPV-based LEDs is electron injection and transport.174,192 Incorporating a

Grimsdale et al.

phenylene-based polyacid between two layers of PPV has been reported to increase the EL efficiency 250 times.193 Heterostructure LEDs made from self-assembled multilayers of PPV and polyanions such as sulfonated polyaniline have been constructed; charge recombination at the interface is presumed to cause emission of bright yellow-green light with low drive voltages.194-202 The emission can be blue-shifted by increasing the thickness of the nonconjugated polyanion layers.203 Polarized emission has been obtained from a stretched PPV film, with emission parallel to the stretching direction being 8 times stronger than that perpendicular to it.204 Rubbing alignment of films of the precursor polymer 3 prior to conversion gives even better anisotropy (12 times stronger emission parallel to the rubbing direction).205,206 Micropatterning of PPV films has been achieved by patterning films of the precursor polymer 3 with a scanning tunnelling microscope (STM) tip prior to conversion.207,208 The areas affected by the STM show no emission, possibly due to oxidation of the precursor. Patterning of PPV has also been achieved by a microlithographic process,209 by holographic lithography,210 by micromolding of the precursor film,211,212 and by screen or inkjet printing.212 Improved efficiencies in PPV devices can also be achieved by the deliberate introduction of disorder into the polymer main-chain, thus interfering with polymer chain packing. Son et al.213 have synthesized amorphous PPV by using a modified Wessling synthesis with a xanthate group. The resulting PPV is claimed to have a mixture of cis- and transvinylene units which reduces conjugation and interferes with packing. Efficiencies of up to 0.22% for single-layer devices with aluminum-aluminum cathodes and 2% for double-layer devices using a dispersion of a diaryloxadiazole in poly(methyl methacrylate) as electron-transporting layer were obtained. Subsequent work with solution-processible PPVs and a well-defined cis/trans microstructure has not substantiated this observation as general.162 A simple modification of the PPV synthesis gives access to PPV-derived blue light-emitting polymers with a high concentration of a fluorescent chromophore yet at the same time having only a short conjugation length. The objective is achieved by partial conversion of precursor homopolymer 3 or 4 to produce statistical copolymers 16 and 5 with regions of saturated and unsaturated units (Scheme 5).214,215 Conversion of 3 to 16 can be achieved either thermally or by treatment with base as mentioned above. Studies on the elimination of PPV precursor polymer 2 have shown that the degree of elimination can be influenced by careful control of conversion conditions.216,217 Incompletely converted PPV is reported to be unstable with devices showing poorer stability, although higher initial efficiency.218 Heeger and co-workers have studied thermal partial conversion and found that conversion at 160 °C for 2 h gave the copolymer 16 with the best stability and emission efficiency.219 As indicated before, polymer 3 undergoes substitution with methanol at 50 °C to give polymer 4, another precursor polymer to PPV. Complete conversion of polymer 4 to PPV (1) can only be achieved by simultaneous heat and acid treatment. Thermal treatment alone results in partial conversion to copolymer 5 with intermittent regions of saturated and unsaturated units. This copolymer can also be obtained by leaving a solution of polymer 3 in methanol at ambient temperature for several days and then thermally converting the resulting copolymer in which the sulfonium groups have been partially displaced by methanol.220 The UV-vis


spectrum of copolymer 5 prepared in this way indicates that the polymer contains conjugated sequences mainly comprising stilbene and distyrylbenzene units. Films of copolymer 5 so formed are thermally stable against further elimination, requiring combined heat and acid to be converted to PPV (1). When incorporated in an EL device using aluminum and ITO as electrodes, copolymer 5 shows a broad emission spectrum and the emitted light appears blue-green to the eye. The EL maximum is distinctly blue-shifted (2.44 eV, λmax ) 508 nm) compared with that of PPV (2.25 eV, λmax ) 551 nm),215,221 indicating that the EL emission of 5 is closer to that reported for distyrylbenzene (2.6 eV, λmax ) 475 nm).222 The device efficiency is twice that of PPV in the same device configuration, which correlates well with the fluorescence yields of the two polymers.214 A similar tuning of the conjugation length of PPV prepared by the Vanderzande sulfinyl precursor route can be achieved by partial oxidation of the sulfinyl groups to sulfones, followed by selective elimination of the former.223,224 CVD copolymerization of a bis-chloromethylbenzene with an unsubstituted cyclophane enables the controlled introduction of unsubstituted ethanediyl units into the resulting PPV chloro precursor, so that the PL maximum of the final PPV can be tuned between 470 and 525 nm. LEDs using this material show blue (λmax ) 420 nm) emission.225 Copolymers of PPV and poly(vinyl alcohol) (PVA) have been prepared by spin coating a mixture of the precursor 4 and PVA, followed by thermal conversion. Partial displacement of the sulfonium groups in 4 by the hydroxyl groups of the PVA gives rise to cross-linked materials with short conjugation lengths. By increasing the ratio of PVA used, the PL emission maximum may be shifted from green (λmax ) 550 nm) to blue (λmax ) 485 nm).226 Partial (7000 h).228 These lifetimes, however, are still well short of those needed for commercially viable devices, and as a result, PPV is no longer being actively examined as a suitable emissive material for such applications.

2.2. Alkoxy-, Thioalkyl-, and Alkylamino-Substituted PPVs Substitution of PPV (1) with alkoxy groups has two effects. First, the band gap is reduced, leading to a marked red-shift in emission. Thus, while PPV shows yellow-green PL with maxima at 2.25 eV (λmax ) 551 nm) and 2.4 eV (λmax ) 520 nm),1 the monomethoxy derivative 18 emits in


the yellow with a single maximum at 550 nm,230,231 and the 2,5-dimethoxy polymer 19 shows orange-red emission with maxima at 2.06 eV (λmax ) 603 nm) and 1.91 eV (λmax ) 650 nm).232 Second, poly(2,5-dialkoxy-1,4-phenylene vinylene)s (RO-PPVs) with at least one long, solubilizing alkoxy side-chain, e.g. poly[2-methoxy-5-(2-ethylhexyl)-1,4phenylene vinylene] (MEH-PPV, 20), dissolve in organic solvents, such as chloroform, toluene, or tetrahydrofuran.233-241 Branched side-chains, as in MEH-PPV, generally enhance solubility more than linear ones with the same number of carbon atoms. In addition, the fluorescence efficiency of polymers bearing longer side-chains is increased by several orders of magnitude compared with that of the dimethoxy polymer 19. Longer alkyl chains also lower the glass transition temperature of the polymers.242

The first LEDs using such soluble PPV derivatives were fabricated independently by researchers at Sumitomo233,234 and at Santa Barbara.235,236,238,240,243 They have become the most widely studied group of electroluminescent conjugated polymers. Such soluble polymers no longer require thermal treatment during device fabrication, which is the drawback of the PPV precursor route. As the sulfonium precursor polymer 2 can be processed from methanolic solutions and like PPV (1) does not dissolve or swell in chloroform, whereas most of the substituted PPVs used in EL devices can be processed from chloroform, one can design multilayer devices consisting of sandwich structures with layers of different polymers, Viz. PPV and substituted PPVs.244,245 Mono- and dialkoxy-substituted PPVs can be prepared by a sulfonium precursor route as described for the synthesis of PPV (1).87,232,240,246-250 Delmotte et al.251 have studied the synthesis of poly(2,5-diethoxy-1,4-phenylene vinylene) (21) by the sulfonium precursor route and found that partial elimination of the sulfonium groups occurred when >1 equiv of base was used. This elimination was found to be selective for short conjugated fragments, yielding only tetraethoxystilbene units even when up to 35% of leaving groups had been eliminated. They explained this as being due to an E2 mechanism where the formation of tetraethoxy stilbene units represents the highest gain in resonance energy at each step. These units gave blue emission centered at an emission maximum of 415 nm. There has been much interest in controlling the conjugation length of alkoxy-PPVs, both in order to achieve the entire color spectrum in LEDs and because the presence of nonconjugated segments on the backbone of dialkoxy-PPV derivatives is reported to increase their EL efficiency.252 A fairly straightforward method for the preparation of 2,5-


Grimsdale et al. Scheme 6. Synthesis of PPV/Dialkoxy-PPV Copolymers: (a) NaOH, MeOH/H2O, 0 °C; (b) HCl; (c) Dialysis (Water); (d) MeOH; (e) 220-300 °C, vacuum; (f) 220 °C, HCl(g)/Ar

dimethoxy substituted PPV 19 with varying degrees of conjugation was presented by Ramakrishnan and coworkers253,254 using a process already developed by Cambridge Display Technology.228 Sulfonium polyelectrolyte precursor polymers prepared Via the standard Wessling route were subjected to competitive nucleophilic substitution using varying ratios of methanol and NaOAc in acetic acid. The acetate groups underwent selective thermal elimination upon heating to 250 °C, affording partially conjugated statistical copolymers 22 showing colors from light yellow to deep red. [We use the extended wavy line connection (as in structure 22) to represent a statistical copolymer.] As the degree of conjugation decreased, the PL and EL efficiencies increased.255 This approach was readily adopted to prepare MEH-PPV precursors 23 but failed in the case of unsubstituted PPV (1).256 Similar control of conjugation length has also been obtained by means of the precursor polymers 24, in which the xanthate groups could be selectively eliminated.257 A similar approach developed by Hwang et al.258 is to replace some of the sulfonium groups with thermally stable thiophenoxy groups prior to thermolysis. The PL maximum of the resulting copolymers 25 has been tuned between 507 and 585 nm. A similar control of the conjugation length of the methoxyPPV 18 prepared by the Wessling method has been obtained by performing dialysis of the sulfonium precursor with

methanol-water (9:1) for varying lengths of time. Thermal conversion gave precursor polymers 26 with varying degrees of nonconjugation. Longer dialysis times produced greater degrees of nonconjugation with a concomitant blue-shift in emission. A double-layer device in which a layer of moreconjugated 26 was deposited upon a layer of less-conjugated 26 showed an unusual emission spectrum with peaks at 560 and 590 nm, and an increase in efficiency over single-layer devices. The peak at 590 nm did not correspond to the emission maximum of either layer and was attributed to a new optical interaction at the interface between the layers. If the order of the layers was reversed, these effects were not observed.259-261 The sulfonium groups in the precursors to alkoxy-PPVs prepared by the Wessling route are more labile toward nucleophilic substitution than those in the PPV precursor 2, as is shown by the tendency for gelation in the preparation of dimethoxy-PPV 19, due to displacement of sulfonium groups by water.232 This has been used to good advantage by the Cambridge group to produce copolymers of PPV and dialkoxy PPV chemically tuned to possess improved properties for EL and for patterned waveguides. This is exemplified by the preparation of copolymers incorporating phenylene vinylene (1) and dimethoxyphenylenevinylene 10 repeat units (Scheme 6), although other copolymers have also been investigated.163,262,263 Statistical copolymerization of the sulfonium monomers 2 and 27 affords a statistical copolymer 28, whose composition is simply controlled by using different feed ratios of the two monomers. The copolymerization solvent mixture permits selective substitution of the tetrahydrothiophene groups by methanol, as this occurs at a faster


rate at those benzylic carbons which are attached to the more activated dimethoxyphenylene rings. The resulting copolymer 29 has two leaving groups with methoxy groups being adjacent to all dimethoxyphenylene and a small amount of phenylene units, and sulfonium groups adjacent to most of the phenylene units During subsequent heat treatment (220 °C for 2 h at 10-5 mbar) of spin-coated thin films (thickness ca. 100 nm) of polymer 29, the sulfonium groups are completely eliminated whereas the methoxy groups mainly remain, to give homogeneous, dense, and uniform films of copolymer 30 which are stable toward further heating. NMR spectra confirm the presence of methoxy leaving groups after thermal conversion, while the IR spectra show a decrease of a diagnostic band for the benzylic methoxy groups. The longest-wavelength absorption band of polymer 30 exhibits a distinct blue-shift with respect to that of polymer 29 which increases as the content of dimethoxyphenylene units in the copolymer rises, showing that the polymer backbone consists of a statistical assembly of discrete conjugated segments separated by saturated units. The EL efficiency of polymer 30 varies with the composition of the copolymer and has its peak with a copolymer obtained from a feed ratio of m/n ) 9:1. The maximum efficiency measured in this case was 0.3%, which was 30 times higher than that of PPV (1) (0.01%) in the same device configuration (Al-Al2O3/polymer/Al).221,262 As both PPV and 30 are insoluble in chloroform, it is possible to make multilayer devices, containing these polymers and also soluble PPV derivatives, with a wide variation in the thicknesses and order of the layers. Such devices allow the color and brightness of the emitted light to be controlled, which clearly emphasizes the scope which copolymers offer as organic light-emitting materials.245 In fact, the basis of this invention lies at the heart of all commercial solutionprocessible light-emitting polymer technology. When the thermal treatment of copolymer 29 is carried out in the presence of hydrogen chloride, the fully converted polymer 31 is obtained. Polymer 31 can also be prepared by thermal elimination of the sulfonium precursor copolymer 29 obtained from Wessling copolymerization under the standard conditions.264-266 The efficiency of LEDs using an emitting layer of 31 is increased 10-fold by use of a PPV holetransporting and a polyquinoline electron-transporting layer.267 A similar copolymer 32 is soluble in organic solvents due to the longer dodecyl chains. The emission maximum of these copolymers can be tuned between the maxima of the homopolymers at 515 and 590 nm, simply by altering the composition.268

The Cambridge group also used this behavior advantageously in a patterning process for copolymer films. Conversion of thin films of copolymer 29 at 220 °C under vacuum produces tetrahydrothiophene and hydrogen chloride as byproduct, which normally escape quickly. In principle, HCl can promote further conversion of polymer 30 to 31, if the acid stays long enough within the copolymer film. It does so in thick copolymer films, or in thin films when the

Chemical Reviews, 2009, Vol. 109, No. 3 907 Scheme 7. Synthesis of MEH-PPV (20): (a) 3-(Bromomethyl)heptane, KOH, EtOH, reflux; (b) HCHO, concentrated HCl, dioxane; (c) KOtBu, THF

byproducts are deliberately trapped. This can be used to produce two colors within a single film, by depositing an aluminum mask onto the precursor copolymer surface according to a predefined lithographic pattern. The aluminum cap provides a physical barrier to the HCl, which causes the formation of more conjugated polymers. After etch-removal of the metal mask, the polymer that was underneath it shows a deeper orange appearance to the eye than the uncovered material and a striking difference in refractive index.269,270 In an EL device with ITO and Ca contacts, such a two-colorpatterned copolymer shows emission from the regions containing the partially conjugated polymer 30.262,263 This observation is consistent with the fact that the partially conjugated material 30 has a higher quantum efficiency for EL than the fully conjugated polymer 31, and it demonstrates the ability to produce passive emissive elements by a negative lithography process. The same patterning procedure can also be applied to the design of waveguides.262,270 The Vanderzande sulfinyl precursor method has also been used to make alkoxy-PPVs.139,223,271,272 The synthesis of the monomers is complicated compared with those for other routes but has recently been greatly improved.273,274 MEH-PPV was first synthesized by Wudl using the Wessling sulfonium precursor route,240 and as shown above, the Wessling and Vanderzande routes may be used to make a variety of mono- or dialkoxy-substituted PPV derivatives (RO-PPVs). However, in most cases an alternative method first developed by Gilch275 is preferred, involving polymerization of 1,4-bis-halomethylated benzene derivatives which are susceptible to base-promoted 1,6-elimination, with a large excess of base, usually KOtBu.233,234,238-240,243,276 This approach for MEH-PPV was first disclosed in the patent literature.241 This shortens the preparation of the conjugated polymer by two steps compared with the Wessling route and consequently usually increases the yield. Scheme 7 shows the synthesis of the most studied dialkoxy-PPV-MEH-PPV (20).240,243 The bis-chloromethyl monomer 33 is synthesized in good yield by alkylation of 4-methoxyphenol 34 with 2-ethylhexyl bromide, followed by chloromethylation.240 Polymerization of the monomer 33 with a 10-fold excess of potassium tert-butoxide in tetrahydrofuran (THF) gives the bright red-orange MEH-PPV (20), which is purified by several reprecipitations from methanol. A problem with this reaction is the formation of insoluble gels of high molecular weight polymer (gelation), but this can be overcome by use of DMF as solvent.277 Benzyl halides can be used as chain terminating agents, either to control the molecular mass (Vide infra) or to introduce terminal functional groups.278 Electroluminescent RO-PPVs other than MEH-PPV have been similarly prepared.233,239,240,272,279-284 The polymerization has also been performed under phase-transfer catalysis conditions,285 reportedly giving better molecular masses and lower polydispersities than the standard method.286 Polym-


erization of a mixture of monomers gives statistical copolymers whose properties are determined by the ratio of monomers.268,281 Molar masses of soluble RO-PPVs, obtained by polymerization of bis-chloromethyl compounds with KOtBu, are comparable with the same polymers prepared by the sulfonium precursor route. Interestingly, comparison of OC1C10-PPV (35a) (where C10 is 3,7-dimethyloctyl) made by the Gilch and Vanderzande methods showed that the former had higher PL and EL efficiencies though the latter was obtained in better yield and had fewer structural defects.272 If 30%). Copolymers 46 with 10-20% of nonconjugated segments have been found to have increased EL efficiencies compared with fully conjugated RO-PPVs.241 The OC1C10-PPV copolymer 46a, in which the chloroprecursor was treated with 2-dimethylaminoethanol, shows much higher efficiencies than the R-methoxy copolymer 46b, with indium (1.1% vs 0.16%) or gold (0.25% vs 0.01%) cathodes after an initial “burn-in” period.384,385


Quaternization of the amines with methyl iodide to give 46c raises the efficiencies for indium and gold cathodes to values (1.3% and 1.1%, respectively) which are equal to that of a device with a calcium cathode (1.2%).384,385 Photoprocessed, micropatterned LEDs have been constructed using copolymer 46c with 5% nonconjugated segments.386 The photoprocessing was found to increase the effective conjugation length concomitant with a drop in PL and EL quantum efficiencies. An alternative method for controlling the conjugation length of ROPPVs developed by Burn is to treat the chloro precursor 45 with tributyltinhydride to replace some of the chloro substituents with hydrogens (Scheme 9).387 The proportion of saturated ethanediyl units in the resulting copolymers 47 after thermal elimination of the remaining chloro substituents corresponds closely to the number of equivalents of the hydride used. In a reverse procedure, treatment of the dihexyloxypolymer 39d with N-halosuccinimdes to reintroduce a proportion (up to 30%) of R-halogens has also been shown to blue-shift the PL by interrupting conjugation and to increase fluorescence efficiency, as the bulky halogen atoms reduce interchain interactions.388

Other routes have also been used to make RO-PPVs. Ho¨rhold and co-workers have prepared MEH-PPV (20) by


the Horner coupling of a bis-aldehyde and a bisphosphonate.389-391 The PL and EL properties of this material are reported to match those of the standard material from the Gilch route, but the solubility is better due to its lower molar mass. This method can also be used to make alternating copolymers.390,391 The copolymers 48a and 49 with dimethoxy-phenylene units show similar properties to those of MEH-PPV 20, but the copolymer 48b with dioctyloxy-phenylene moieties shows much more long wavelength emission, possibly due to greater aggregation.391

The copolymer 50 shows orange (λmax ) 585 nm) EL emission in forward, reverse, and alternating current (ac) modes.392 Dimethoxy-PPV (19) has been made by vapor deposition polymerization of a bischloromethyl monomer, and thermal elimination of the resulting chloro precursor polymer.393 Low molecular mass polymers can be prepared by Stille coupling of dialkoxyaryl dihalides or ditriflates with bis-stannylethene.394,395 The cyclophane-substituted PPV 51 prepared by Stille coupling shows yellow-green EL (λmax ) 563 nm).396 Cross metathesis of 2,5-dialkoxy-1,4-divinylbenzenes produces oligomeric 39e-g.397-399 Cathodic electropolymerization of 2,5-bis(dibromomethyl)decyloxybenzene has been used to make the decyloxy-PPV 52.400 Lahti and co-workers have used Heck coupling of a 4-vinylbromobenzene AB monomer to make regioregular (head-to-tail) poly(2-dodecyloxyphenylene vinylene) (53) (Scheme 10), which shows strong yellow PL and EL (λmax ) 542 and 540 nm, respectively).401 So far there has been no comparative study on regiorandom vs regioregular dodecyloxy-PPV 53, but it is known that regioregularity in other conjugated polymers, e.g. poly(3-alkylthiophene)s (see section 3.6 below), can lead to better electrical and optical properties.

Scheme 10. Synthesis of Regioregular Poly(2-dodecyloxyphenylene vinylene) (53): (a) Br2, HOAc; (b) MeOH, H+; (c) C12H25Br, K2CO3; (d) LiAlH4; (e) PCC; (f) Ph3PCH2Br, BuLi; (g) Pd(OAc)2, P(o-tolyl)3, DMF


Yoshino and co-workers402 have reported that poly[(2,5dinonyloxy-1,4-phenylene) vinylene] (39h) forms a nematic liquid crystalline phase upon melting. Cooling of the melted material to room temperature gives an oriented film with a lower energy gap (2.08 eV, 597 nm) than for an unoriented film (2.21 eV, 562 nm) formed by spin-casting of a chloroform solution of the polymer. Rubbing of a film of this material is reported to induce molecular orientation, giving rise to polarized electroluminescence.403 Attempts have also been made with varying success to induce liquidcrystallinity (and thus potentially polarized emission) in ROPPVs by attaching side-chains bearing mesogenic groups. Thus, the polymer 54 forms a smectic liquid crystalline phase

Grimsdale et al.

and shows green-yellow PL (λmax ) 550 nm).404 The EL is bias-dependent with broad yellow (λmax ) 600 nm) emission in forward bias but blue (λmax ) 400 nm) emission from the mesogen in reverse bias. A similar liquid crystalline polymer 39i made by Heck coupling405 shows only weak EL (λmax ) ca. 560 nm) and unpolarized PL.406 The polymer 55 and copolymer 56 with azobenzene mesogens in the side-chains have been reported to show yellow-green PL (λmax ) 550 nm), but there is no published data on liquid crystalline properties or polarized emission.407,408 Introduction of chiral substituents can lead to circularly polarized luminescence. Meijer and co-workers have produced circularly polarized EL from devices using the copolymer 57.409 The best results were obtained for a copolymer with an m:n ratio of 8:1. Holmes and co-workers have developed synthetic routes toward 2,3-dialkoxy-substituted PPVs 58.410-416 Their more recent improved route412,415 is shown in Scheme 11. By comparison with the classical 2,5-substituted PPV derivatives, this substitution pattern results in a notable blue-shift in the emission maximum accompanied by an increase in the solidstate PL efficiency. As a result, the dibutoxy-PPV 58a shows green emission (λmax ) 530 nm) with a PL efficiency of 40%410,416,417 (cf. 15-20% for MEH-PPV235,418). The efficiencies of single-layer devices (λmax ) 520 nm) using 58a were up to six times higher than those for similar devices using PPV.417,419 Efficiencies were even higher for the double-layer devices ITO/PPV/58a/Ca. The EL efficiency can also be boosted up to 200 times by use of electron-transporting layers.420 Increasing the steric constraints by using branched chains as in 58b leads to a further blue-shift in the emission (λmax ) 513 nm), accompanied by a drop in PL efficiency to 28%.413-415 LEDs using polymer 58b (λmax ) 505 nm) show only slightly higher efficiency than PPV devices.413-415 Polymers have been prepared with one methoxy tris(ethoxy) (59) (λmax ) 601 nm) or two methoxy mono-, bis-, and trisethoxy- side-chains (60) [λmax ) 640 nm].421-425 The disubstituted polymers with short side-chains (60a,b) are best made by the Wessling route,425 but the Gilch route has been found to be superior for those with longer side-chainss59 and 60c.422-425 These polymers show very low EL efficiency in LEDs due to their low (

Synthesis of Light-Emitting Conjugated Polymers for Applications in

Recommend Documents